Variation in the composition of eastern Australian molasses was examined. Data on molasses composition for 1997–2001 were collected from all 28 eastern Australian sugar mills. Since the last detailed study in 1975, there have been decreases in the concentration of reducing sugars (–6.2%), sucrose (–0.9%), total sugars (–2.5%), other organic matter (–11%) and calculated metabolisable energy (ME) (–2.8%). Dry matter (DM, +0.1%) and ash (+29%) concentrations have increased. Differences between milling regions were observed for DM (P<0.001), reducing sugars (P<0.05) and sucrose (P<0.01) concentrations. Molasses DM was highest for the Northern milling region and lowest for the Southern milling region. Reducing sugar concentrations were lower whereas sucrose concentrations were higher in molasses from the Southern milling region than the Northern region. The concentration of reducing sugars, total sugars and calculated ME was significantly (P<0.05) lower in mid-season molasses than early and late season molasses. Mid-season molasses ash concentration was highest (P<0.05). Significant (P<0.05) region × year interactions were found for all molasses components. Significant (P<0.05) region × season interactions were found for reducing sugars, sucrose, total sugars, ash and the calculated ME. Significant (P<0.001) nested effects for mills within a region were found for all molasses components. Mill within-region effects (P<0.001) accounted for much (31–62%) of the total variation for all molasses components. Significant (P<0.01) nested effects for season within year were found for all molasses components. Depending on the molasses component, season within year accounted for 3–19% of variation. Calculated ME in molasses was most correlated with total sugars (r = –0.97; P<0.001) and ash concentration (r = –0.73; P<0.001). Commercial cane sugar (percentage sugar in juice) was negatively (r = –0.43; P<0.001) correlated with calculated molasses ME. Generally, the observed variation in the calculated ME of molasses would only be sufficient to reduce average daily gains by up to 0.05 kg/day in cattle fed diets containing 60% molasses. Compared with the calculated molasses ME of molasses in 1975, a similar reduction in animal performance could be expected with current molasses composition. Such variation in ME also has implications for the use of molasses in the dairy industry and for ethanol production.